Natalizumab’s airplane mode: PML vs. cessation relapses


The implementation of risk mitigation strategies in medicine is highly dependent on the context, and there is a thin line between an unwanted adverse event being considered as a medical error or an unfortunate portion of bad luck. Similar kinds of ambiguous risk assessments are implemented in our daily life. Many of our friends and colleagues have been injured by car accidents and almost never we dig into the traffic design factors that facilitate their occurrence. However, when an airplane crashes extensive investigations need to elucidate the underlying mechanism and preferably find a human or technical culprit. 


One of the domains in which extensive risk mitigation strategies are in place is the prevention of progressive multifocal leukoencephalopathy (PML) caused by the JC virus when being treated with natalizumab (Tysabri). PML implies progressive inflammation of the brain’s white matter and can be fatal. The risk of PML in the most ‘risky’ situation (read: highest JC antibody index > 1.5 ) is 1 in 5,000 in the first 12 months of treatment, 1 in 1,111 in the first 24 months, and 1 in 333 in the first 36 months. The risk never exceeds 1 in 100. Based on the results of the NOVA trial and other observational cohorts, the PML risk is reduced with 88% when switching to 6-weekly natalizumab infusions instead of the current standard practice of 4-weekly infusion intervals. 

Based on these risks, it’s considered good practice by many neurologists to determine the JC-index before offering natalizumab and exclude pwMS from using this drug if they carry the virus regardless of the JC-antibody index. This is in sharp contrast with the risks of relapses and disability accrual we tolerate for female pwMS who fall pregnant under natalizumab and their newborns. 

In case of a pregnancy while being treated with natalizumab, experts in the field recommend to continue natalizumab until the end of the second trimester and restart the drug asap after giving birth (preferably within 4 weeks). The rationale for this recommendation is the significant risk of rebound disease activity and thus disability when stopping natalizumab, the known increased risk of relapses up until three months after giving birth and the wish to reduce fetal exposure to natalizumab during the third trimester. During this period, natalizumab is actively transported over the placenta. In practice, many female pwMS will put more weight on the uncertainties regarding natalizumab’s effects on the foetus and decide to quit natalizumab during the first term or even before. 

Based on current registries, natalizumab exposure during pregnancy does not appear to increase the risk of spontaneous abortions and congenital malformations. However, exposure during the third trimester is associated with mildly/moderately reduced platelet and red blood cell counts (which are known to resolve spontaneously). One small study including 17 infants with natalizumab exposure during pregnancy also showed a lower birth weight and higher risk of hospitalisation in the first year of life. 

In a recent German cohort study, Hellwig et al. observed the risk and impact of rebound relapses in people who fall pregnant while being treated with natalizumab. They collected data on 274 pregnancies in 251 women. The following numbers are noteworthy: 

  • The majority of participants stopped natalizumab early during the first trimester (n = 189, 68.97%), and 85 participants (31.02%) stopped prior to pregnancy. 
  • Relapses were reported in 183 pregnancies (66.78%), severe relapses in 44 pregnancies (16.05%), and potentially life-threatening relapses in 3 pregnancies (1.10%). These three women became all bedbound by their relapses, and 2 now live in nursing homes. 
  • 10% of women retained clinically meaningful disability from pregnancy-related natalizumab cessation relapses 1 year after giving birth. 
  • There was no protective effect of pregnancy on relapse risk in the third trimester. 
  • Neither exclusive breastfeeding nor restarting natalizumab within 4 weeks after giving birth were associated with a reduced risk of time to first post-delivery relapse 6 months after delivery (although there was a trend when restarting natalizumab early).

Although many pwMS and clinicians go in “airplane mode” when hearing about the risk of PML associated with natalizumab use, the risk for female pwMS of becoming disabled because of a natalizumab cessation relapse during or after pregnancy is manifold higher. If you fall pregnant, stop natalizumab infusions during the first trimester and only start with them again after completing breastfeeding without having had a relapse, you have gambled and been exceptionally lucky. Personally, I feel this risk of disabling relapses after stopping natalizumab is unacceptable and undoes all the efforts we make to treat pwMS early in their disease course. In addition, the MS expert advice should be updated and include the recommendation to 1) share these raw data with female pwMS,  2) include the option to continue natalizumab throughout the entire pregnancy and 3) implement a 6-week extended-dosing interval during pregnancy to minimise fetal exposure. 

Twitter: @SmetsIde

Disclaimer: Please note that the opinions expressed here are those of dr. Ide Smets and do not necessarily reflect the position of the Barts and The London School of Medicine and Dentistry nor Barts Health NHS Trust.

JAMA Netw Open 2022 Jan 4;5(1):e2144750. doi: 10.1001/jamanetworkopen.2021.44750.

Multiple Sclerosis Disease Activity and Disability Following Discontinuation of Natalizumab for Pregnancy

Kerstin Hellwig, MD; Marianne Tokic, MSc; Sandra Thiel, PhD; Nina Esters, MD; Charlotte Spicher, MD; Nina Timmesfeld, PhD; Andrea I. Ciplea, PhD; Ralf Gold, MD; Annette Langer-Gould, MD, PhD

  • PMID: 35072719
  • DOI: 10.1001/jamanetworkopen.2021.44750

Importance: The magnitude of risk of pregnancy-related multiple sclerosis relapses, particularly severe relapses, following natalizumab cessation is unclear, as is whether this risk is reduced by pregnancy or other modifiable factors. Objective: To determine the association of early natalizumab withdrawal before or during pregnancy with risk of severe relapses and relapse-related disability. Design, setting, and participants: This prospective cohort study used data from the German Multiple Sclerosis and Pregnancy Registry, which enrolled participants between November 2006 and February 2018. Data were collected through structured telephone-administered questionnaires and review of neurologists’ notes. Registry patients who stopped natalizumab treatment within the 2 years before or in the first trimester of pregnancy were included in this analysis. Data were analyzed between January and November 2021.Exposures: Cessation of natalizumab before pregnancy or until the first trimester. Main outcomes and measures: Severe and significant relapse-related disability was defined as at least a 2.0-point increase on the expanded disability status scale or new or worsening relapse-related ambulatory impairment. Multivariable models accounting for measures of disease severity and repeated events were used. Results: The cohort comprised 255 women with 274 pregnancies (mean [SD] age at pregnancy onset, 31.25 [4.27] years) who stopped natalizumab before pregnancy (n = 85; median time before last menstrual period, 14.29 weeks [IQR, 3.14-42.43 weeks]) or in the first trimester (n = 189). During pregnancy and the postpartum year, relapses were reported in 183 pregnancies (66.78%), severe relapses in 44 pregnancies (16.05%), and potentially life-threatening relapses in 3 pregnancies (1.10%). One year post partum, significant relapse-related disability was accrued in 29 pregnancies (10.58%). Relapses during pregnancy (n = 109; 39.78%) and in the postpartum period (n = 135; 49.27%) were common. Pregnancy (as a time-dependent covariate) was not associated with a reduced relapse risk (adjusted HR, 0.90; 95% CI, 0.64-1.27). Neither exclusive breastfeeding (adjusted HR, 1.34; 95% CI, 0.86-2.10) nor restarting natalizumab within 4 weeks post partum (adjusted HR, 1.06; 95% CI, 0.48-2.36) were associated with a reduced risk of early postpartum relapses 6 months after delivery. However, the relapse rate ratio during 12 months post partum was lower (0.49; 95% CI, 0.28-0.86) when natalizumab was restarted in the first 4 weeks after birth. Conclusions and relevance: This cohort study’s finding suggest that 10% of women may retain clinically meaningful disability from pregnancy-related natalizumab cessation relapses 1 year post partum. This information should be shared with women on natalizumab who desire pregnancy to weigh the high risk of pregnancy-related relapses and disability to the partly uncertain risks of continuing natalizumab throughout pregnancy or switching to depleting agents before conception.

About the author

Ide Smets


  • Thanks for sharing. I think the largest PR problem for Tysabri is that outdated and conditional PML risk numbers are tossed aroud very casually between Patient treating Neurologist without sufficient understanding of context. It appears that everyone always fixates on numbers no matter how inaccurate they in fact are.

    Wouldn’t it therefore be a good if we ourselves started to better qualify the risks estimated in 2014 instead of stating matter-of-fact like ” The risk of PML with JC antibody index > 1.5 is 1 in 5,000 in the first 12 months of treatment, 1 in 1,111 in the first 24 months, and 1 in 333 in the first 36 months” Are these really still the actual risks that someones faces when starting treatment in 2022?

    – Where do we account for improved monitoring guidelines and a much broader range of immunosupperessants that patients could have received before starting Tysabri? Isn’t there a need for an update?
    – What about the improved treatment methods that have been put in place after discovery of a PML? Is the 20% mortality rate still realistic?

    • Hi Marc,

      Thank you for this comment, very sharp analysis about what we still need to know to improve our risk assessments.

      Regarding the official PML risk numbers, you are right. They are certainly partially outdated as the Biogen cohort dates from 2014. I have seen other anecdotical reports come by with PML cases reported with indexes below 0.9 etc, but none of them really allows to update these numbers as big cohorts are needed to say anything reliable about a low-incidence event such as PML. The current numbers stem for 249 PML cases out of an initial cohort of over 30,000 pwMS. Nonetheless, we could definitely benefit from more granularity in the low JC index pwMS as they are the ones now put on NTZ and there is clearly a difference in pwMS with index below 0.6 and between 0.6 end 0.9.

      The treatment for PML has not evolved over the preceding decade, and it’s only the improved detection/monitoring that has reduced mortality. Some centres use PLEX when a diagnosis of PML is made, but there’s no proven efficacy. Survival mainly depends on whether PML was detected in its asymptomatic phase and potentially also by the extent to which IC support is available in a specific country. In asymptomatic PML patients mortality is likely to be lower than 20%.

  • Hi Ide. Nothing to do with MS, but your comment in respect to car accidents. I am not sure if you are aware but In the UK we have a system call Road Safety Audits. Once a new scheme has been constructed any reportable accident (to the police) is reviewed at what is called a RSA stage 4. This specifically looks at the accidents (and the cause of them) and if there is any contributing factor from the road design. Regards

By Ide Smets



Recent Posts

Recent Comments